The No.1 Website for Pro Audio
 Search This Thread  Search This Forum  Search Reviews  Search Gear Database  Search Gear for sale  Search Gearslutz Go Advanced
Lossy vs Lossless encoded music, do you hear the difference? Dynamics Processors (HW)
Old 2nd June 2014
  #61
Gear Head
 

Quote:
Originally Posted by JimK View Post
I consider myself somewhat of an audio snob and I can definitely hear the difference between a 192/24 FLAC file and an MP3. :-)

Jim
Would you be so kind to prove it by taking this blind test (see signature)? iTunes quality (AAC 256k) is used instead of mp3.
__________________
cdvsmp3.wordpress.com
Old 2nd June 2014
  #62
Gear Head
 

Quote:
Originally Posted by walter88 View Post
The play tests work ok in Safari and Opera on Windows now (both CD and HD test), but there are no file links on the test pages in either. Nothing clickable except the radio buttons and checkboxes.
If you check the box "Show me the download link", all online links are hidden because you are supposed to play the music contained in the downloaded zip file.
You say the PLAY link works now. If it does, make sure you do not check the box "Show me the download link", as this would hide the links to the online audio files as explained.
__________________
cdvsmp3.wordpress.com
Old 2nd June 2014
  #63
Gear Head
 

Quote:
Originally Posted by walter88 View Post
The play tests work ok in Safari and Opera on Windows now (both CD and HD test), but there are no file links on the test pages in either. Nothing clickable except the radio buttons and checkboxes.
There was indeed an issue with the form in Opera. I think it is solved now. Let me know if it isn't. Sorry!
__________________
cdvsmp3.wordpress.com
Old 3rd June 2014
  #64
Lives for gear
 
RRCHON's Avatar
I tired to complete your A/B tests but the samples are too short and the popping in between them is too painful to hear over and over again.

It's not really testing my ability to differentiate two audio different digital audio formats, that's more of a test of if I can stay sane while doing it. Sorry, if you make a kinder gentler version where I can actually hear more than switching 6 second clips, I would be willing to try again.
Old 3rd June 2014
  #65
Lives for gear
 
Viruslabs's Avatar
 

I know one of these public mp3 versus wav tests where after closer investigation it turned out that the software used had engaged the computer os internal samplerate conversion ! nice one..

So be carefull what you do such tests with..

Also always nice to use productions that have no lowend and highend.. so really >40 db down at 40 and 15 khz already.. So better check the testfiles with a spectrum analyzer before wasting time with them.

Also your speaker system has to be able to show a weakness in the low department.. so down to 30hz flat.
And dont tell me that music is only played on systems that cant go that low..
We dont live in the gramophone age anymore but in the time where 21 inch woofers are more popular than ever before thanks to 4000 watts amps that adequately drive them.. When you want to make sure that a data compression dont has any audible losses you need to use the best speaker system as refference. That mp3 is good enough for the kitchen radio is known..
But is it as good as losslees..and is lossless lossless? you only find out when listening to good speakers and at a hot level

And than beware the type of test music that dont contains any high quality sound sources at all.. 100% itb and voice with 200% autotune treatment... I ve witnessed productions like this to actually starting to sound better by mp3 encoding with a lower rate.. There are tracks out there that just sound a little smoother on 96kps than on 256 kbs,..

But there are also classical recordings that are just stringy mids close up that pretty well survive even low coding rates.

In general sometimes mp3 just works..as better and cleaner the source material as better the mp3 yo get from it. So you need to test it with various music and also not highpolished audio.

Actually the best test material i ve came along was transfers from good sounding vinyl records..
That gives the mp3 encoding a hard time because of all the fine noise and rumble in the recording.

Is pretty easy to find music that converts to mp3 beautifully.. when you want to show that mp3 has no significant loss that is possible. But that the virtual reality of that moment and dont speaks for any possible audiomaterial. You should be aware that the data reduced file is permanently damaged and an inferior copy of the original recording.
Thats just the fact.
Even the Fraunhofer institute themself dont sell their mp3 as lossless..they only claim that most people dont get it under normal listening conditions.
But what are normal listening conditions and who is most people?

Consequently the only thing you can prove with such self tests is whether you are one of most people or not.
Old 3rd June 2014
  #66
Lives for gear
 

lol
Old 3rd June 2014
  #67
Gear Head
 

Quote:
Originally Posted by RRCHON View Post
I tired to complete your A/B tests but the samples are too short and the popping in between them is too painful to hear over and over again.

It's not really testing my ability to differentiate two audio different digital audio formats, that's more of a test of if I can stay sane while doing it. Sorry, if you make a kinder gentler version where I can actually hear more than switching 6 second clips, I would be willing to try again.
I am sorry the track format is unpleasant enough to prevent you from completing the test. I think the layout I've chosen serves the purpose, and I have had quite a bit of positive feedback so far, but I will consider some alternative method to include people like you. Thank you very much for trying!
__________________
cdvsmp3.wordpress.com
Old 3rd June 2014
  #68
Lives for gear
 
RRCHON's Avatar
Thanks for considering it.

Just eliminating the pops between switches and sections a bit longer would make it useful for me. Older people have slower brains and it takes me at least 20 seconds to even begin to hear things critically, but the pops and clicks in between just reset my hearing so they are to me at least - the worst offenders.

Let me know if you change it, I'll be happy to go through the process.
Old 4th June 2014
  #69
Quote:
Originally Posted by bogosort View Post
Like JimK's post before yours, this is a meaningless statement until you specify the type of mp3 and the source material. All mp3 encoding is not equal.

If you believe that the differences between 320 kbps mp3 and lossless is glaringly obvious on most material, then I'd recommend some blind listening tests.
Even a 16 Bit .wav file to mp3 320kbs presents an obvious quality loss to anyone who knows what to listen for. Blurred transients, lack of bass definition, a mufukin wool blanket covering the speakers, I dono...
Old 4th June 2014
  #70
Lives for gear
 
bogosort's Avatar
Quote:
Originally Posted by chronmaster View Post
Even a 16 Bit .wav file to mp3 320kbs presents an obvious quality loss to anyone who knows what to listen for. Blurred transients, lack of bass definition, a mufukin wool blanket covering the speakers, I dono...
Have you tested this assertion blind?
Old 4th June 2014
  #71
Lives for gear
 

Quote:
Originally Posted by chronmaster View Post
Even a 16 Bit .wav file to mp3 320kbs presents an obvious quality loss to anyone who knows what to listen for. Blurred transients, lack of bass definition, a mufukin wool blanket covering the speakers, I dono...
I used to have a blanket like that, got rid of it, too many holes.
Old 4th June 2014
  #72
Gear Head
 

Quote:
Originally Posted by walter88 View Post
Forgot to ask. Someone else who took the test asked what their results were. Are results not disclosed to the tester?
Sorry for not answering before...
I reveal the score, but not the solution key to allow more than one try, which would be good for the level of confidence of the result. I would if I were able to randomize the sequence, but I can't due to limitations in the Adobe Forms service. I am working on some improvements... based on the many suggestions I am getting from the forum and testers. I look forward to yours. I'll look for an alternative filesharing service to host downloadable files.

Thanks for the interest!

__________________
cdvsmp3.wordpress.com
Old 4th June 2014
  #73
Quote:
Originally Posted by chronmaster View Post
Even a 16 Bit .wav file to mp3 320kbs presents an obvious quality loss to anyone who knows what to listen for. Blurred transients, lack of bass definition, a mufukin wool blanket covering the speakers, I dono...
Quote:
Originally Posted by bogosort View Post
Have you tested this assertion blind?
There are some files (typically associated with electronic music/heavy studio FX and sounds not just not found in nature but not found in free air) that are reportedly known to be 'codec killers' where the codecs choke on the content and produce anomalies. But, aside from those, there doesn't seem to be much/any evidence of even trained listeners being able to consistently tell 320's from full CD-A that I've come across.

Although, of course, personal, anecdotal claims that there are 'obvious' differences are frequent.

But those seem to generally evaporate when properly tested.
Old 4th June 2014
  #74
Gear Addict
 

Quote:
Originally Posted by chronmaster View Post
Even a 16 Bit .wav file to mp3 320kbs presents an obvious quality loss to anyone who knows what to listen for. Blurred transients, lack of bass definition, a mufukin wool blanket covering the speakers, I dono...
No offense chronmaster, but that's what everybody says before they try it blind (me included). But then nobody can do it blind. The brain's a funny thing.

Last edited by walter88; 5th June 2014 at 07:37 PM.. Reason: added blind
Old 5th June 2014
  #75
Lives for gear
 
Viruslabs's Avatar
 

Quote:
Originally Posted by theblue1 View Post

But those seem to generally evaporate when properly tested.
the only proper test i know is to listen to such files on a system that does at least 30hz to 17 khz flat at a sound pressure of around 105-115 db c
And than you let the dj play the 2 tracks in sync thru the samplerate conversion of its pioner cdj player and he just moves the crossfader between the mp3 and the wav.. will there be an audible difference? The crazy people that have done such unrepresentative test that just show what that cd player makes out of it..what of cause has nothing to do with what another codec would do to that file..and an additional samplerate conversion just for a little pitching is also no proper treatment of an mp3 file.. But.. I just described the real world application where mp3 hurt the most..

So it really dont matters what your living room abx test seem to prove to you..When stage reality does sound different.
And thats no superstition or myth its just like that.. By now i use mp3s to tune pa´s just to make sure that the highs get smooth enough and the lowend a little extra. But i personally think that dj´s that use mp3´s should´nt be booked at all because using such low quality formats is disrespectful towards the audience.
And disrespectful towards the producers of that music since the used mp3´s are mostly "free" copies. So low quality files and low quality moral go hand in hand..as smaller the file as easier it is to give your mate 200 tracks from last week via bluetooth while having a beer.. Saying this i actually think that pono is a very good initiative and it should be forbiden to play low resolution files on public events..making 192k uncompressed files mandatory just to ensure that only professional and legaly purchased files can be played on big public events..
Old 5th June 2014
  #76
Lives for gear
 
RRCHON's Avatar
Quote:
Originally Posted by Viruslabs View Post
the only proper test i know is to listen to such files on a system that does at least 30hz to 17 khz flat at a sound pressure of around 105-115 db c
That's a pretty big hurdle on frequency response, the cost to setup a test environment, loudspeakers and room that would preform at those specifications isn't chump change.

For most people that kind of environment is just not relevant and 100dBSPL isn't comfortable especially with hyper-compressed music.

Also after years of abuse my ears would ring for days listening to anything that loud for more than 5 minutes. I really don't ever see a reason to exceed 80-85 decibels of sound pressure and even that is loud.

I'm all for lossless audio as I don't really think the storage or transmissions hurdles are what they once were and therefore no longer relevant for music purchases (streaming still would benefit from data compression,) but your standards are unrealistic and irrelevant for the majority of people who consume audio on those media.
Old 5th June 2014
  #77
Gear Head
 
TheDrumschlag's Avatar
 

I am 100% confident that I can tell the difference between lossy and lossless using a good studio monitoring chain. It's less specific qualities that I could point out and more just the overall feeling of the track. There's a sort of "openness" that is really apparent in lossless files in comparison to lossy. I can't hear a huge difference between 44.1 and 96k tho, and I think anything above 96 is pretty much just wasting disk space.
Old 5th June 2014
  #78
Gear Addict
 

Quote:
Originally Posted by TheDrumschlag View Post
I am 100% confident that I can tell the difference between lossy and lossless using a good studio monitoring chain. It's less specific qualities that I could point out and more just the overall feeling of the track. There's a sort of "openness" that is really apparent in lossless files in comparison to lossy. I can't hear a huge difference between 44.1 and 96k tho, and I think anything above 96 is pretty much just wasting disk space.
You should try it. Blind. It can be very enlightening. Or confusing.
Old 5th June 2014
  #79
Gear Addict
 

Seriously, if anyone can tell 320kbps lossy consistently blind, you should contact the trustmeimascientist guy and beat his challenge, because no one has yet. The Boston Audio Society and AES tests showed people had been fooling themselves about 44.1khz vs 96khz, and then 320k lossy MP3 was later added to the mix in internet challenges and people still couldn't tell. I'm sure someone somewhere can tell 320k lossy every time simply because the files are not what the originals were, they're measurably different in the audible range, and the lossy processes themselves are taking a lot for granted. And yet no one who can do it consistently has proven that. Blind is very different from sighted, on any system.

Last edited by walter88; 5th June 2014 at 09:39 PM.. Reason: added in the audible range
Old 6th June 2014
  #80
Lives for gear
 
Viruslabs's Avatar
 

blind is difficult but a/b on a big system works even with closed eyes and fingers in the ear because you feel the difference in the bass allready.. uncompressed is more open and goes deeper since the compression also sees very low fequencys as redundant information it can reduce.. so it reduces them. Might not work with any track but we did such a test lately and made jokes about internet guys that insist on mp3 datareduction beeing in audible.. was a good laugh watching a guy with fingers in the ears getting it right each time..
Old 6th June 2014
  #81
Gear Addict
 

I'd be interested to know which codecs affect the bass, since low frequencies are the easiest to encode.
Old 6th June 2014
  #82
Yes, mainly from the high frequencies.
Old 6th June 2014
  #83
okok let me do the test
Old 6th June 2014
  #84
Gear Head
 

Hi there,
I've made major changes in the approach of my CD vs AAC 256k test (visit the URL in the signature), based on suggestions from some of you in the forum and from testers. I would be very grateful if you tried it and provided feedback. Thank you!
__________________
cdvsmp3.wordpress.com
Old 6th June 2014
  #85
Gear Head
 

Quote:
Originally Posted by walter88 View Post
Seriously, if anyone can tell 320kbps lossy consistently blind, you should contact the trustmeimascientist guy and beat his challenge, because no one has yet. The Boston Audio Society and AES tests showed people had been fooling themselves about 44.1khz vs 96khz, and then 320k lossy MP3 was later added to the mix in internet challenges and people still couldn't tell. I'm sure someone somewhere can tell 320k lossy every time simply because the files are not what the originals were, they're measurably different in the audible range, and the lossy processes themselves are taking a lot for granted. And yet no one who can do it consistently has proven that. Blind is very different from sighted, on any system.
I totally agree... and blind is the only way to go to prove anything on the subject. Prove what you claim for.
But What about iTunes plus (AAC 256k VBR)? I consider myself to have a pretty good sense of hearing, and to know where to look for differences when trying a blind test... and I have done quite well up to mp3 192k CBR, but I am consistently defeated when trying AAC 256k. I have no reason to think that my ears are exceptional, though, so I wonder if anyone of you can tell CD and AAC 256k apart. I've devised an online blind test that you guys might be interested in trying. It is not the ideal ABX testing tool, but it is as close as I can get for now.
__________________
cdvsmp3.wordpress.com
Old 8th June 2014
  #86
Great test dude. I pulled about 50%. I intuitively felt there was more headroom on the CD quality clips, but after going through the 16 clips a second time, I lost objectivity.
Old 8th June 2014
  #87
Gear Head
 

Quote:
Originally Posted by chronmaster View Post
Great test dude. I pulled about 50%. I intuitively felt there was more headroom on the CD quality clips, but after going through the 16 clips a second time, I lost objectivity.
Thanks! Were you able to submit your results? (maybe it is just that you didn't use you gearslutz.com nickname...)
(Ok, I see them now, coming from China...)

__________________
cdvsmp3.wordpress.com
Old 8th June 2014
  #88
Gear Addict
 

Quote:
Originally Posted by Viruslabs View Post
blind is difficult but a/b on a big system works even with closed eyes and fingers in the ear because you feel the difference in the bass allready..
Well, we'd concluded in an earlier post that you had probably irreparably compromised your hearing ability from all that extended listening at 105 - 115dB. Now we learn that you don't actually use your ears anyway - preferring to block your ears with a finger and "feel" the sound.

So let's hope you don't cheat on that blind test by sneaking a peek at the extent of the excursions on your trouser-flapping. :-)
Old 8th June 2014
  #89
Lives for gear
 
Viruslabs's Avatar
 

Quote:
Originally Posted by Tom McC View Post
Well, we'd concluded in an earlier post that you had probably irreparably compromised your hearing ability from all that extended listening at 105 - 115dB. Now we learn that you don't actually use your ears anyway - preferring to block your ears with a finger and "feel" the sound.

So let's hope you don't cheat on that blind test by sneaking a peek at the extent of the excursions on your trouser-flapping. :-)
You seem to cheat reason anyway.. so your problems is beyond listening. beside the thread question was regarding lossless compression..
Not about lossy compression..so this is all ot..

But its funny that nobody questions that lossy compression is lossy..except you guys..
not even the frauenhofer institute themself claims mp3 to be lossless And the also admit that the illusion only works with normal listening levels.. and normal for them is less than your avarage studio monitoring level.. ..

Mp3 dont measures lossless..and its also audible not lossless.. You can fabricate as many test that seem to show otherwise as you want..that wont change these facts.

Maybe you should engage other senses when your ears fail you so badly and dont question mine....therefor i suggested the blind test even deff people can experience..and ..it does make a difference.. just try it..
Old 8th June 2014
  #90
Lives for gear
 
Viruslabs's Avatar
 

Quote:
Originally Posted by chronmaster View Post
Great test dude. I pulled about 50%. I intuitively felt there was more headroom on the CD quality clips, but after going through the 16 clips a second time, I lost objectivity.
thats the magic of an abx test.. do it a 3rd time and you will swear that everything sounds the same..

beside..pulling 50% right in such a statistical method will be accounted as no audible difference since it could be explained with randomn choices.. So when the well made mp3 just manages to cheat you often enough it allready leads to a false conclusion , since you have heard the more openess too..aslong your ears wasnt tired, as so many others..
And since we hear that openess statement so often we should conclude statisticly that there is indeed something that gets experienced as more openess.. you hear that again ..and again and again..and maybe thats the less stressive quality that got discovered lately.
Maybe it just feels more open for your less stressed brain.. and as soon the abx test has led to an allways stressed condition you dont get a difference anymore.. would fit..

But still people rather buy that oppeness to be an illusion than that the 10 times smaler file might having no losses to be the illusion?
Something that stresses the listener more seems to have more contradictonary information that needs to be processed.. Thats a loss of quality.
Just is that audible or only receptable? And does that difference really matter? aslong it feels worse it is worse. The weakness in the bass is also not really audible since that happens below 20 hz..what has an fx on the transmisson of high energy sound systems.. but not on your home stereo.. So audible or not.. there are fx..and you can feel them.

I am fine with the claim that highrates do a good job an beeing allmost inaudible on normal listening levels..thats fine..and not more does the frauenhofer institute claim. But going further than them and claiming it to have no losses what soever is just
Post Reply

Welcome to the Gearslutz Pro Audio Community!

Registration benefits include:
  • The ability to reply to and create new discussions
  • Access to members-only giveaways & competitions
  • Interact with VIP industry experts in our guest Q&As
  • Access to members-only sub forum discussions
  • Access to members-only Chat Room
  • Get INSTANT ACCESS to the world's best private pro audio Classifieds for only USD $20/year
  • Promote your eBay auctions and Reverb.com listings for free
  • Remove this message!
You need an account to post a reply. Create a username and password below and an account will be created and your post entered.


 
 
Slide to join now Processing…
Thread Tools
Search this Thread
Search this Thread:

Advanced Search
Similar Threads
Thread
Thread Starter / Forum
Replies
[email protected] / So much gear, so little time
3
feyshay / Music Computers
8
pikapikapi / So much gear, so little time
2
asylum / High end
8
runamuck / So much gear, so little time
8

Forum Jump
Forum Jump